Proposal #8

	This proposal should not be considered for the best proposal prize.

	I would give this proposal a 9/10 for overall clarity and organization. It's a well written paper, but lacks details.

	The overall idea is ambitious without overdoing it. I do have concerns about the "learning" involved in this project. 7/10

	The proposal details how the researcher will program a robot to identify zippers on clothing using the tactile and visual senses of the robot. This research will be useful in
robotics, if robots are ever to work alongside humans.

	The proposal is very clear and avoids most technical jargon. It meets most of the proposal guidelines, however it is a little too short at only 8 pages (not counting the title
page or the references). It is obviously a step forward in the field of Developmental Robotics, as well as being an impressive achievement in its own right.

	As far as related work goes, the researcher has done a good job of including relevant references. It appears to be a rather extensive list.

	My favorite part of this project is that it is both cool and useful. The identification and use of zippers are  obviously something that's necessary should robots work in human
environments given zippers (and related fasteners) are everywhere.

	The part of the proposal that I dislike is that it seems more like a computer vision project than a developmental robotics one. Where is the artificial intelligence in this
project? It looks as though the robot may be able to recognize zippers using vision alone (there are numerous references to a "visual model"). I wish the researcher would have
detailed the project a bit more thoroughly, and included a "Possible Expansion" section with AI capabilities.

	I am curious to know if the zipper types and sizes will be varied. Also, a lot of zippers are covered by flaps, like on coats and blue jeans (as mentioned in the proposal). How
do you intend to deal with this?

	Something that may be of assistance to the robot is to track the type of objects (blue jeans, jacket, backpack, etc.) and track where the zipper is most commonly found on each
object. For example you are most likely to find a zipper running in a long straight line down the middle of a jacket. On a backpack the zipper curves over the top of the backpack.
If the robot remembers the type of object it would certainly speed up the search. On the other hand, this is not an object classification problem and this suggestion probably
belongs in the future work section.

	The research team certainly has the capabilities and resources necessary to complete the project. No problems there.

	This is a decent project and a decent proposal. I just wish there was a bit more detail in the proposal. Good luck!